Today we are excited to share Granite 4.0 Nano, our smallest models yet, released as part of IBM's Granite 4.0 model family. Designed for the edge and on-device applications, these models demonstrate excellent performance for their size and represent IBM's continued commitment to develop powerful, useful, models that don't require hundreds of billions of parameters to get the job done.

Like all Granite 4.0 models, the Nano models are released under an Apache 2.0 license with native architecture support on popular runtimes like vLLM, llama.cpp, and MLX. The models were trained with the same improved training methodologies, pipelines, and over 15T tokens of training data developed for the original Granite 4.0 models. This release includes variants benefiting from the Granite 4.0’s new, efficient hybrid architecture, and like all Granite language models, the Granite 4.0 Nano models also carry with them IBM's ISO 42001 certification for responsible model development, giving users added confidence that models are built and governed to global standards.

Specifically, Granite 4.0 Nano comprises of 4 instruct models and their base model counterparts:

Building sub-billion to ~1 billion parameter models is an active and competitive space, with advancements in performance and architectures recently made by a number of model developers such as Alibaba (Qwen), LiquidAI (LFM), Google (Gemma) and others. 0 Nano models demonstrate a significant increase in capabilities that can be achieved with a minimal parameter footprint, as measured by a series of general benchmarks across General Knowledge, Math, Code, and Safety domains.

Source link